The Five-Step Funnel Audit I Run in My First Week at Any Startup
Before running a single new campaign, I map the funnel and find the leak. Here is the exact process, with real numbers from one audit.
When a startup brings me in to fix growth, the founders usually want new channels. Paid social, influencer deals, a podcast tour. I understand the impulse. But in almost every case the fastest gains come from fixing what already exists. So my first week is always the same audit.
Here it is, step by step, with numbers from a recent engagement with a meal-kit subscription startup.
Step 1: Write the funnel down in plain words
I list every stage a customer passes through, using the company's actual product, not a generic template. For the meal-kit business, it looked like this:
- Visits the homepage
- Views the menu or plans page
- Starts checkout
- Enters a delivery postcode
- Completes payment
- Receives the first box
- Orders a second box
Steps 4 and 6 are often missing from standard funnels, and that is exactly where trouble hides.
Step 2: Get one month of counts for each stage
Not percentages yet, raw counts. For one month, the numbers were roughly 42,000 visits, 18,500 plan views, 3,900 checkout starts, 3,100 postcode entries, 1,050 payments, 1,010 first boxes, and 540 second orders.
Step 3: Calculate stage-to-stage conversion
This is where the story appears. Visit to plan view was 44 percent, which is healthy. Postcode entry to payment was only 34 percent, which jumped out immediately. And first box to second order was 53 percent.
Step 4: Watch real sessions at the worst stage
I spent two hours watching session recordings of people who entered a postcode and then left. A clear pattern emerged: shipping costs and delivery days only appeared after the postcode, and many visitors saw a delivery day that did not suit them. They left without a word.
Step 5: Size the opportunity before proposing fixes
If we lifted postcode-to-payment from 34 to 45 percent, that would mean roughly 340 extra paying customers a month with no additional traffic. For comparison, the founders' proposed paid social campaign was forecast to bring around 200 new customers a month at significant cost.
What we did
We showed delivery cost and available days on the plans page, before checkout. We also added a second delivery day for the two busiest postcodes. Within six weeks, postcode-to-payment rose to 41 percent. Not the full target, but a meaningful, cheap improvement.
Why this works
- Fixing leaks compounds. Every future visitor benefits.
- It builds shared understanding. Once the team saw the funnel written down, product and marketing stopped arguing about whose fault growth was.
- It makes later spending smarter. Paid traffic poured into a leaky funnel is expensive.
New channels have their place. But in my experience, the first and cheapest growth is usually hiding in a stage nobody bothered to measure.
Cover: Photo by Kaleidico on Unsplash.
